Data Transmission Method

Abstract

To avoid packet collisions at an aggregation station even if the number of sensors that transmit data increases, thereby preventing the occurrence of packet loss.SOLUTION: There is provided a data transmission method in which, in a data transmission target area 10 where a plurality of sensors 30 are installed, observation data from each sensor 30 is assigned to an index corresponding to a frame so as to be transmitted to an aggregation station 40 by a PLIM method, the frame being formed of a plurality of time slots and a use channel. The data transmission method executes each of: advance preparation step (S1) of aggregating, as a sensor observation data occurrence probability distribution, an observation probability for the observation data at each sensor 30 and the position of an observation target; step (S2) of using the sensor observation data occurrence probability distribution based on the observation data to calculate the index having a minimum cumulative error between packet data that can be received at the aggregation station 40 by the PLIM method and replacement data when the packet data cannot be received; and step (S3) of transmitting the observation data to the aggregation station 40 by the PLIM method to which the index is applied.SELECTED DRAWING: Figure 4

Description

[Technical Field]

[0001] The present invention relates to a data transmission method. [Background technology]

[0002] In recent years, the Internet of Things (IoT) has become widespread. In the IoT, attention has been focused on Low Power Wide Area (LPWA), a communication method that enables long-distance communication with low power consumption. In order to resolve the decrease in throughput in LPWA, some of the inventors of this application have proposed, in Patent Document 1 (JP 2022-85522 A) and Non-Patent Document 1, etc., a data transmission method using a packet-based index modulation method (so-called Packet Level Index Modulation (PLIM) method) that sets the time slot for transmitting packets and the channel to be used based on the information bit sequence to be transmitted. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-85522 [Non-patent literature]

[0004] [Non-Patent Document 1] National University Corporation, University of Electro-Communications, "New communication method to solve IoT issues - Increased data transmission volume with low power consumption - │," [online], July 9, 2021, National University Corporation, University of Electro-Communications, [Retrieved June 22, 2022], Internet<URL:https: / / www.uec.ac.jp / news / announcement / 2021 / 20210709_3546.html> Summary of the Invention [Problem to be solved by the invention]

[0005] In the data transmission methods disclosed in Patent Document 1 and Non-Patent Document 1, as shown in Figure 13, when a sensor receives radio waves from a radio wave source and transmits data to an aggregate station, an index having multiple index numbers is created based on the time slot and channel used by the sensor. The data transmitted by the sensor to the aggregate station is then associated with the index, thereby improving data transmission throughput. However, as the number of sensors increases, it becomes more likely that multiple sensors will simultaneously transmit the same index number to the aggregate station, as shown in Figure 14, which increases the likelihood of so-called packet collisions. When such packet collisions occur, it becomes impossible for the aggregate station to decode the data (packet loss).

[0006] Therefore, attempts have been made to avoid packet collisions at the aggregation station by randomly assigning index numbers to each sensor, which correspond to the data that the sensor transmits to the aggregation station.However, because the index numbers are simply assigned randomly, the success rate of avoiding packet collisions at the aggregation station varies, and as the number of sensors increases, packet collisions at the aggregation station cannot be avoided, so the above-mentioned problem remains unresolved. [Means for solving the problem]

[0007] Therefore, the present invention aims to prevent the loss of data that should be received at the aggregation station, even if a packet collision occurs at the aggregation station and packet loss occurs when the number of sensors transmitting data increases, by mathematically optimizing the assignment of data to be transmitted and index numbers in the index used in the PLIM method.

[0008] That is, the present invention provides a PLIM (Packet Level Index Multimedia Information System) that, in a data transmission area where a plurality of sensors are installed to observe an observation target, packetizes observation data from each of the sensors, assigns the packetized observation data to one of indexes corresponding to frames formed by a plurality of time slots and channels used, and transmits the packetized observation data. a step of calculating, by using the sensor observation data occurrence probability distribution based on the observation data from each of the sensors within the data transmission target area, the index that minimizes a cumulative total of errors between packet data received at the aggregate station in the PLIM method used by each of the sensors and replacement data when the packet data cannot be received at the aggregate station, and a step of transmitting the observation data from each of the sensors to the aggregate station using the PLIM method to which the index has been applied, wherein the observation data is quantized by a plurality of observation data numbers corresponding to each required numerical range, and the step of calculating, by using the index that minimizes a cumulative total of errors between packet data received at the aggregate station in the PLIM method used by each of the sensors and replacement data when the packet data cannot be received at the aggregate station, is performed based on the following equation:

number

[0009] This allows mathematical optimization of the assignment of observation data to be transmitted and index numbers, so that even if a packet collision occurs at the aggregation station when the number of sensors transmitting data increases, replacement data can be used in place of data that could not be received, preventing the loss of data that should be received at the aggregation station.

[0010] Furthermore, if the observation data is the average value of the observation data at each of the sensors calculated in the preliminary preparation process, it is preferable that each of the sensors omits transmitting the packet data for the observation data to the central station, and that the central station adopts the average value as the replacement data when the packet data cannot be received.

[0011] This makes it possible to prevent loss of data that should be received at the central station, and also to reduce the amount of packet data transmitted (data communication volume) from the sensor to the central station.

[0012] In addition, in a data transmission area where a plurality of sensors are installed to observe an observation target, the observation data from each of the sensors is packetized, and the packetized observation data is assigned to one of indexes corresponding to frames formed by a plurality of time slots and channels used, and transmitted. a step of calculating, by using the sensor observation data occurrence probability distribution based on the observation data from each of the sensors within the data transmission target area, the index that minimizes a cumulative total of errors between packet data received at the aggregate station in the PLIM method used by each of the sensors and replacement data when the packet data cannot be received at the aggregate station, and a step of transmitting the observation data from each of the sensors to the aggregate station using the PLIM method to which the index has been applied, wherein the observation data is quantized by a plurality of observation data numbers corresponding to each required numerical range, and the step of calculating, by using the index that minimizes a cumulative total of errors between packet data received at the aggregate station in the PLIM method used by each of the sensors and replacement data when the packet data cannot be received at the aggregate station, is performed based on the following equation:

number

[0013] This allows mathematical optimization of the assignment of observation data to be transmitted and index numbers, so that even if a packet collision occurs at the aggregation station when the number of sensors transmitting data increases, replacement data can be used in place of data that could not be received, preventing the loss of data that should be received at the aggregation station.

[0014] Furthermore, if the observation data is the most frequent value of the observation data at each of the sensors calculated in the preliminary preparation process, it is preferable that each of the sensors omits transmitting the packet data for the observation data to the central station, and that the central station adopts the most frequent value as the replacement data when the packet data cannot be received.

[0015] This makes it possible to prevent loss of data that should be received at the central station, and also to reduce the amount of packet data transmitted (data communication volume) from the sensor to the central station.

[0016] Furthermore, the advance preparation step preferably comprises dividing the data transmission target area into a plurality of divided areas, and aggregating, for each of the divided areas, the observation probabilities for the numerical values ​​of the observation data at each of the sensors in a specific one of the divided areas as a divided area-specific sensor observation data occurrence probability distribution, and calculating the index that minimizes the cumulative total of errors between the packet data received at the aggregation station in the PLIM method used by each of the sensors and the replacement data when the packet data cannot be received at the aggregation station, by selecting the divided area-specific sensor observation data occurrence probability distribution that is most similar from the divided area-specific sensor observation data occurrence probability distributions, based on the observation data by each of the sensors in the data transmission target area, and calculating the index that minimizes the cumulative total of errors between the packet data received at the aggregation station in the PLIM method used by each of the sensors and the replacement data when the packet data cannot be received at the aggregation station, by using the selected divided area-specific sensor observation data occurrence probability distribution.

[0017] This makes it possible to assign more detailed observation data to be transmitted and index numbers according to the conditions of the observation area. [Effects of the Invention]

[0018] According to the configuration of the present invention, by mathematically optimizing the assignment of observation data to be transmitted and index numbers, even if a packet collision occurs at the aggregation station when the number of sensors transmitting data increases, replacement data can be used in place of data that could not be received, thereby preventing the loss of data that should be received at the aggregation station. [Brief explanation of the drawings]

[0019] [Figure 1] FIG. 1 is a schematic configuration diagram of a data transmission system according to a reference embodiment. [Figure 2] FIG. 10 is a flowchart showing an example of an implementation procedure of a data transmission method according to a reference embodiment. [Figure 3] 10 is an explanatory diagram of mapping data obtained by converting a histogram showing quantized data and frequencies of RSSI values ​​of each sensor obtained in the advance preparation step in the reference embodiment into a Gaussian distribution. FIG. [Figure 4] FIG. 3 is a flowchart showing an example of an implementation procedure of a data transmission method in the first embodiment. [Figure 5] FIG. 10 is a flowchart showing an example of an implementation procedure of a data transmission method according to the second embodiment. [Figure 6] These are the parameters for a comparative simulation of the root mean square error between the number of sensors when transmitting data in a data transmission system and the observation data obtained by using replacement data when packet collisions occur at the aggregation station. [Figure 7] 7 is an explanation of the legend for the comparative simulation performed with the specifications shown in FIG. 6. [Figure 8] The results of the comparative simulations are shown in Figs. 6 and 7. [Figure 9] FIG. 11 is an explanatory diagram of mapping data obtained by converting a histogram of the occurrence probability of RSSI values ​​in each sensor into a Gaussian distribution when the radio wave source is located in the first divided area in the advance preparation step of the third embodiment. [Figure 10] FIG. 11 is an explanatory diagram of mapping data obtained by converting a histogram of the occurrence probability of RSSI values ​​in each sensor into a Gaussian distribution when the radio wave source is located in the second divided area in the advance preparation step of the third embodiment. [Figure 11] FIG. 11 is an explanatory diagram of mapping data obtained by converting a histogram of the occurrence probability of RSSI values ​​in each sensor into a Gaussian distribution when the radio wave source is located in the third divided area in the advance preparation step of the third embodiment. [Figure 12] FIG. 11 is an explanatory diagram of mapping data obtained by converting a histogram of the occurrence probability of RSSI values ​​in each sensor into a Gaussian distribution when the radio wave source is located in the fourth divided area in the advance preparation step of the third embodiment. [Figure 13] This is the first explanatory diagram of the prior art. [Figure 14] This is the second explanatory diagram of the prior art. DETAILED DESCRIPTION OF THE INVENTION

[0020] Hereinafter, an embodiment of a data transmission method according to the present invention will be described with reference to the drawings. Note that in the following embodiment, the data transmission method will be described based on a data transmission system 50 in which the observation target is a radio wave source 20, but the present invention is not limited to this embodiment.

[0021] (Reference embodiment) A data transmission system 50 in this reference embodiment includes one or more radio wave sources 20 in a data transmission target area 10, multiple sensors 30 that receive radio waves emitted from the radio wave sources 20, and an aggregation station 40 that aggregates observation data transmitted from the multiple sensors 30. The multiple sensors 30 in this reference embodiment quantize the signal strength (hereinafter referred to as RSSI value) of the radio waves received as observation data, assign the quantized data to an index number in the PLIM method, and transmit the data to the aggregation station 40. One of the features of this reference embodiment is that the allocation of index numbers used by the multiple sensors 30 is optimized to minimize conflicts in the aggregation station 40 between index numbers transmitted from the multiple sensors 30.

[0022] As shown in Fig. 1, the data transmission system 50 in this reference embodiment includes at least one radio wave source 20, a first sensor 32, a second sensor 34, a third sensor 36, and a fourth sensor 38 in a data transmission target area 10, and an aggregation station 40 installed at one location. Hereinafter, the first sensor 32, the second sensor 34, the third sensor 36, and the fourth sensor 38 may be collectively referred to as "sensors 30." Note that the number of sensors 30 installed in the data transmission system 50 is not limited to the number shown in this reference embodiment. Each step will be described below with reference to the flow diagram shown in Fig. 2.

[0023] In order to optimize the allocation of index numbers used by each sensor 30, a preparation step (S1) is performed to compile quantized data of RSSI values ​​and the distribution of their occurrence frequency (sensor observation data occurrence probability distribution: hereinafter referred to as mapping data 60) for each sensor 30 when each radio wave source 20 transmits radio waves within the data transmission target area 10. By providing such a preparation step, it is possible to analyze the frequency of use of quantized data of RSSI values ​​for each sensor 30 (reception probability (observation probability) for the received signal strength of observation data received by the sensor 30) depending on the geographical conditions, etc., of the data transmission target area 10.

[0024] FIG. 3 is an explanatory diagram of mapping data 60 obtained by converting quantized data (RSSI numbers: observation data numbers) of RSSI values ​​and histograms showing frequency counts for the first sensor 32, the second sensor 34, the third sensor 36, and the fourth sensor 38, obtained in the advance preparation step, into a Gaussian distribution. While the mapping data 60 in FIG. 3 has been converted into a Gaussian distribution, it may also be converted into another probability distribution or left as a histogram. The number of events used in creating FIG. 3 is not particularly limited, but to ensure reliability, it is preferable that the number of events be 100 or more (case 100 or more). By collecting the distribution of the frequency of occurrence of RSSI numbers for each sensor 30 in this way, it is possible to understand the tendency of RSSI number selection by each sensor 30 in the data transmission target area 10.

[0025] Next, a step (S2) of optimizing the design of indexes used in the PLIM scheme used in the data transmission target area 10 in this reference embodiment is performed using the mapping data 60 shown in FIG. 3 . Specifically, this step calculates the index that minimizes the probability that the RSSI values ​​likely to occur in each of the first sensor 32 to the fourth sensor 38 will have the same index number (probability of packet collisions occurring in the aggregation station 40). To achieve this, in this reference embodiment, the design is performed as a problem of minimizing collisions between index numbers of quantized data of RSSI values ​​of two sensors. This allows the target of optimization design to be expressed as a quadratic model, making it possible to create a calculation formula for a quadratic integer programming problem such as the following equation. The calculation formula for such a quadratic integer programming problem can be solved by a known method. Specifically, a mathematical optimization solver such as the Gurobi Optimizer can be used.

number

[0026] Constraint 1 in the above equation states that, for all i and all j, the i-th sensor must assign the j-th observation data number (here, the number of the quantized data of the RSSI value: hereinafter referred to as the quantized data number) to one of the index numbers. Similarly, Constraint 2 states that, for all i and all k, the i-th sensor must assign at most one observation data number (the quantized data number of the RSSI value) to the k-th index number. Similarly, Constraint 3 states that the maximum number of duplicated indexes is the number of sensors. Constraints 1 to 3 make it possible to prevent the calculation of meaningless solutions (where all elements are 0) to the quadratic integer programming problem equation.

[0027] By solving the mapping data 60 shown in Fig. 3 as the quadratic integer programming problem shown in the above equation, it is possible to calculate an appropriate combination (that minimizes packet collisions) of the quantized data numbers of the RSSI values ​​and index numbers (numbers indicating positions within a frame separated by time slots and channels in use) for each sensor 30. A step (S3) of transmitting data in a PLIM system using indexes having index numbers calculated (designed) in this way is executed. The quantized data of the RSSI values ​​of the observation data transmitted from each sensor 30 in this way (hereinafter referred to as transmitted data) is aggregated in the aggregation station 40 using the index numbers, and a step (S4) of decoding the data in a predetermined procedure is executed.

[0028] Even when the index numbers obtained as described above are used, packet collisions (duplicate index numbers) may occur in the aggregation station 40. When a packet collision occurs in the aggregation station 40, the aggregation station 40 executes a process of replacing at least one piece of transmission data where the packet collision occurred with replacement data in order to prevent a loss of data that should be received by the aggregation station 40. In this reference embodiment, quantized data of an RSSI value randomly selected within a preset numerical range is used as the replacement data. Note that the replacement data may also be quantized data of an RSSI value randomly selected within the numerical range of observation data actually observed in the advance preparation step.

[0029] (First embodiment) The basic configuration of the data transmission system 50 in this embodiment is common to the data transmission system 50 described in the reference embodiment, and therefore, the same components as those in the reference embodiment are assigned the same reference numerals as those in the reference embodiment, and redundant explanations will be omitted here. Below, each step in the data transmission method of this embodiment will be explained with reference to the flow diagram shown in Figure 4.

[0030] In designing the index numbers to be used by each sensor 30, a preparation step (S1) is performed to compile quantized data of RSSI values ​​from each sensor 30 when each radio wave source 20 transmits radio waves within the data transmission target area 10, and the distribution of their occurrence frequencies (sensor observation data occurrence probability distribution: hereinafter referred to as mapping data 60). The preparation step (S1) is the same as that in the reference embodiment. In this embodiment, an observation data average value calculation step (S2) is further performed to calculate the average values ​​of the RSSI values ​​of the observation data from the first sensor 32, the second sensor 34, the third sensor 36, and the fourth sensor 38 based on the data collected in the preparation step (S1).

[0031] Next, a step (S3) of optimizing the design of indexes used in the PLIM method used in the data transmission target area 10 in this embodiment is executed using the mapping data 60 shown in Fig. 3. Specifically, this is a step of calculating an index that can minimize the cumulative total of errors (root mean square errors) between the transmission data that should be received and the replacement data when replacement data is used to prevent loss of data that should be received at the aggregation station 40 when the same index numbers corresponding to the quantized data numbers of the RSSI values ​​in the observation data of each of the first sensor 32 to the fourth sensor 38 overlap (when a packet collision occurs at the aggregation station 40).

[0032] To achieve this, in this embodiment, when the index numbers of the quantized data numbers (transmission data) of the RSSI values ​​of two sensors overlap (packet collision) and the aggregation station 40 is unable to receive data, the average value of the RSSI values ​​of the observed data is used as replacement data. This minimizes the cumulative error (root mean square error) between the transmission data that the aggregation station 40 should have received and the replacement data. This allows the target of optimization design to be expressed as a quadratic model, making it possible to create a calculation formula for a quadratic integer programming problem such as the following equation. The calculation formula for such a quadratic integer programming problem can be obtained using a mathematical optimization solver such as Gurobi Optimizer, as in the reference embodiment.

number

[0033] Constraints 1 to 3 in Equation 4 have the same constraint content as in the reference embodiment, and can prevent the calculation of a meaningless solution (all elements are 0) to the equation for the quadratic integer programming problem.

[0034] 3 as the quadratic integer programming problem shown in the above equation, it is possible to calculate an optimal combination of quantized data numbers and index numbers (numbers indicating positions within a frame separated by time slots and channels in use) of RSSI values ​​for each sensor 30 that can minimize the cumulative total of errors (root mean square errors) between the transmission data that should be received at aggregation station 40 and the average value of the RSSI values. A step (S4) of transmitting data in the PLIM method using indexes having index numbers calculated (designed) in this way is then executed.

[0035] The aggregate station 40 then determines whether or not there is a packet collision for the packetized transmission data transmitted from each sensor 30 to the aggregate station 40 (S5). If there is no packet collision (No), the aggregate station 40 aggregates the transmission data as is, and executes a step (S7) of decoding the data using a predetermined procedure. If there is a packet collision at the aggregate station 40 (Yes), the aggregate station 40 uses the average RSSI value of the observation data as replacement data for at least one transmission data item where a packet collision has occurred. That is, the aggregate station 40 substitutes the average RSSI value of the observation data for at least one transmission data item where a packet collision has occurred (S6), and then executes a step (S7) of decoding the data at the aggregate station 40.

[0036] According to the data transmission method of this embodiment, index numbers can be assigned based on the reception characteristics of the RSSI values, which are the observation data of each sensor 30 in the data transmission target area 10. Furthermore, when transmission data from each sensor 30 in the data transmission method using the PLIM method collide at the aggregation station 40, data loss at the aggregation station 40 can be prevented by using the average value of the RSSI values ​​of the observation data calculated in the advance preparation step (S1) as replacement data. Furthermore, the index calculated in this embodiment is designed to minimize the error between the transmission data that should have been received and the average value of the observation data that is the replacement data. Therefore, when a packet collision occurs, the average value of the observation data (average value of the RSSI values ​​of the observation data) is used instead of the transmission data (packet data) that could not be received at the aggregation station 40, thereby preventing loss of transmission data that should have been received at the aggregation station 40. Furthermore, the reliability of decoding transmission data that should have been received at the aggregation station 40 can be improved.

[0037] Here, all observation data from each sensor 30 is converted into quantized RSSI data packets, and the index number corresponding to this packet data is transmitted to the central station 40 as transmission data. However, this is not a limitation. If the RSSI value or quantized RSSI data of the observation data from each sensor 30 is identical to the average RSSI value of the observation data calculated in the preliminary preparation step (S1), data transmission from the sensor 30 that observed the average RSSI value of the observation data calculated in the preliminary preparation step (S1) to the central station 40 can be omitted. In this case, the central station 40 will be unable to receive transmission data from some sensors 30, but it can also execute a process in which the average value of the observation data calculated in the preliminary preparation step (S1) is regarded as transmission data and treated as replacement data. This is advantageous in that the amount of data communication from each sensor 30 can be reduced while maintaining the decoding accuracy of the data to be received by the central station 40.

[0038] (Second embodiment) The basic configuration of the data transmission system 50 in this embodiment is common to the data transmission system 50 described in the reference embodiment and the first embodiment, and therefore, the same components as those in the reference embodiment and the first embodiment are assigned the same reference numerals as those in the reference embodiment and the first embodiment, and redundant explanations will be omitted here. Below, each step in the data transmission method of this embodiment will be explained with reference to the flow diagram shown in FIG.

[0039] In designing the index numbers to be used by each sensor 30, a preparation step (S1) is performed to compile quantized data of the RSSI values ​​of each sensor 30 when each radio wave source 20 transmits radio waves within the data transmission target area 10, and the distribution of the frequency of occurrence (sensor observation data occurrence probability distribution: hereinafter referred to as mapping data 60). The preparation step (S1) is the same as in the reference embodiment and the first embodiment. In this embodiment, an observation data mode calculation step (S2) is further performed to calculate the mode of the RSSI values ​​of the observation data from the first sensor 32, the second sensor 34, the third sensor 36, and the fourth sensor 38, based on the data collected in the preparation step (S1).

[0040] Next, a step (S3) of optimizing the design of indexes used in the PLIM method used in the data transmission target area 10 in this embodiment is executed using the mapping data 60 shown in Fig. 3. Specifically, this is a step of calculating an index that can minimize the cumulative total of errors (root mean square errors) between the transmission data that should be received and the replacement data when replacement data is used to prevent loss of data that should be received at the aggregation station 40 when index numbers corresponding to quantized data numbers of RSSI values ​​in the observation data of each of the first sensor 32 to fourth sensor 38 overlap (when packet collision occurs at the aggregation station 40).

[0041] To achieve this, in this embodiment, when the index numbers of the quantized data numbers (transmission data) of the RSSI values ​​of two sensors overlap (packet collision) and the aggregation station 40 is unable to receive data, the mode of the RSSI values ​​of the observation data is used as replacement data. This minimizes the cumulative error (root mean square error) between the transmission data that the aggregation station 40 should have received and the replacement data. This allows the target of optimization design to be expressed as a quadratic model, making it possible to create a calculation formula for a quadratic integer programming problem such as the following equation. As in the reference embodiment and the first embodiment, the calculation formula for such a quadratic integer programming problem can be obtained using a mathematical optimization solver such as Gurobi Optimizer.

number

[0042] Constraints 1 to 3 in the above equations have the same constraint content as those in the reference embodiment and the first embodiment, and can prevent the calculation of meaningless solutions (where all elements are 0) to the equations for the quadratic integer programming problem.

[0043] 3 as the quadratic integer programming problem shown in the above equation, it is possible to calculate an optimal combination of quantized data numbers and index numbers (numbers indicating positions within a frame separated by time slots and channels in use) of RSSI values ​​for each sensor 30 that can minimize the cumulative total of errors (root mean square errors) between the transmission data that should be received by the aggregation station 40 and the mode value of the RSSI values. A step (S4) of transmitting data in the PLIM method using indexes having index numbers calculated (designed) in this way is then executed.

[0044] The aggregate station 40 determines whether or not there is a packet collision for the RSSI value data (packet data) transmitted from each sensor 30 in this manner (S5). If there is no packet collision (No), the aggregate station 40 aggregates the transmitted data as is, and executes a step (S7) of decoding the data using a predetermined procedure. If there is a packet collision at the aggregate station 40 (Yes), the aggregate station 40 uses the most frequent RSSI value of the observation data as replacement data for at least one packet data for which a packet collision has occurred. That is, the aggregate station 40 substitutes the most frequent RSSI value of the observation data for at least one packet data for which a packet collision has occurred (S6), and then executes a step (S7) of decoding the data at the aggregate station 40.

[0045] According to the data transmission method of this embodiment, index numbers can be assigned based on the reception characteristics of the RSSI values, which are observation data of each sensor 30 in the data transmission target area 10. Furthermore, when packet collision occurs at the aggregation station 40 for packet data transmitted from each sensor 30 by the data transmission method using the PLIM method, data loss at the aggregation station 40 can be prevented by using the mode value of the RSSI values ​​of the observation data calculated in the advance preparation step (S1) as replacement data. Furthermore, the index calculated in this embodiment is designed to minimize the error between the transmission data that should have been received and the mode value of the observation data, which is the replacement data. Therefore, when a packet collision occurs, the mode value of the observation data (the mode value of the RSSI values ​​of the observation data) is used instead of the transmission data (packet data) that could not be received at the aggregation station 40, thereby preventing data loss that should have been received at the aggregation station 40. Furthermore, the reliability of decoding data that should have been received at the aggregation station 40 can be improved.

[0046] Here, all observation data from each sensor 30 is packetized as quantized RSSI data, and the index number corresponding to this packet data is transmitted to the aggregation station 40 as transmission data. However, this is not a limitation. If the RSSI value or quantized RSSI value data of the observation data from each sensor 30 is identical to the mode of the RSSI value of the observation data calculated in the preparation step (S1), data transmission from the sensor 30 that observed the mode of the RSSI value of the observation data calculated in the preparation step (S1) to the aggregation station 40 can be omitted. In this case, the aggregation station 40 will be unable to receive transmission data from some sensors 30, but it can also perform processing to treat the mode of the observation data calculated in the preparation step (S1) as replacement data and treat it as transmission data. This is advantageous in that the amount of data communication from each sensor 30 can be reduced while maintaining the decoding accuracy of the data to be received by the aggregation station 40.

[0047] The applicant conducted a comparative simulation to confirm the error between the original transmission data and the replaced data for each number of sensors when data is transmitted from each sensor 30 to one aggregation station 40 using the indexes calculated in the above embodiment. The applicant conducted the comparative simulation using the simulation specifications shown in FIG. 6 and the simulation legend shown in FIG. 7. FIG. 8 is a graph showing the results of the comparative simulation. When the error in the RSSI number, which is the vertical axis in FIG. 8, is 1, the data value differs by the value of the class width. From this, it can be said that it is preferable for the error in the RSSI number to be less than 0.5.

[0048] As shown in FIG. 8, when the number of sensors is 4, only the all-sensor identical mapping fails to keep the RSSI number error below 0.5. Furthermore, when the number of sensors is 5, the collision probability minimization mapping, the first embodiment, and the second embodiment all achieve an RSSI number error below 0.5. When the number of sensors is 6 or more, none of the legends achieves an RSSI number error below 0.5. Also as shown in FIG. 8, the first and second embodiments perform well among the legends regardless of the number of sensors. Furthermore, it can be said that the second embodiment has the best performance among the legends presented here.

[0049] (Third embodiment) In this embodiment, a step is provided for further detailing the mapping data 60 (reception probability for each piece of observation data) obtained in the advance preparation step in the reference embodiment, the first embodiment, and the second embodiment. That is, as shown in Figures 9 to 12, it is also possible to adopt a configuration in which the data transmission target area 10 is divided into a plurality of divided areas, and the occurrence probability of the RSSI value of each sensor 30 is tallied for each divided area in which the radio wave source 20 is located. Here, a configuration will be described in which the data transmission target area 10 is divided into four divided areas, and one sensor (first sensor 32 to fourth sensor 38) is disposed in each divided area.

[0050] 9 is an explanatory diagram of first mapping data 62 obtained by converting a histogram of the occurrence probability of RSSI values ​​for each sensor 30 into a Gaussian distribution when the radio wave source 20 is located in the first divided area 12. From the first mapping data 62 shown in FIG. 9, when the radio wave source 20 is located in the first divided area 12, the RSSI value obtained by the first sensor 32 located in the first divided area 12 (closest to the first divided area 12) often occurs as a high numerical value. Furthermore, the RSSI value obtained by the fourth sensor 38 located in the fourth divided area 18, which is farthest from the first divided area 12, often occurs as a low numerical value. Furthermore, it can be said that the RSSI values ​​obtained by the second sensor 34 and the third sensor 36 located in the second divided area 14 and the third divided area 16, which are intermediate positions between the first divided area 12 and the fourth divided area 18, often occur as intermediate values ​​between the RSSI value obtained by the first sensor 32 and the RSSI value obtained by the fourth sensor 38.

[0051] 10 is an explanatory diagram of second mapping data 64 obtained by converting a histogram of the occurrence probability of RSSI values ​​for each sensor 30 into a Gaussian distribution when the radio wave source 20 is located in the second divided area 14. From the second mapping data 64 shown in FIG. 10, when the radio wave source 20 is located in the second divided area 14, the RSSI value obtained by the second sensor 34 located in the second divided area 14 (closest to the second divided area 14) often occurs as a high numerical value. Furthermore, the RSSI value obtained by the third sensor 36 located in the third divided area 16, which is the farthest from the second divided area 14, often occurs as a low numerical value. Furthermore, it can be said that the RSSI values ​​obtained by the first sensor 32 and the fourth sensor 38 located in the first divided area 12 and the fourth divided area 18, which are intermediate between the second divided area 14 and the third divided area 16, often occur as intermediate values ​​between the RSSI value obtained by the second sensor 34 and the RSSI value obtained by the third sensor 36.

[0052] 11 is an explanatory diagram of third mapping data 66 obtained by converting a histogram of the occurrence probability of RSSI values ​​for each sensor 30 into a Gaussian distribution when the radio wave source 20 is located in the third divided area 16. From the third mapping data 66 shown in FIG. 11, when the radio wave source 20 is located in the third divided area 16, the RSSI value obtained by the third sensor 36 located in the third divided area 16 (closest to the third divided area 16) often occurs as a high numerical value. Furthermore, the RSSI value obtained by the second sensor 34 located in the second divided area 14, which is farthest from the third divided area 16, often occurs as a low numerical value. Furthermore, it can be said that the RSSI values ​​obtained by the first sensor 32 and the fourth sensor 38 located in the first divided area 12 and the fourth divided area 18, which are intermediate positions between the third divided area 16 and the second divided area 14, often occur as intermediate values ​​between the RSSI value obtained by the second sensor 34 and the RSSI value obtained by the third sensor 36.

[0053] 12 is an explanatory diagram of fourth mapping data 68 obtained by converting a histogram of the occurrence probability of RSSI values ​​for each sensor 30 into a Gaussian distribution when the radio wave source 20 is located in the fourth divided area 18. From the fourth mapping data 68 shown in FIG. 12, it can be seen that when the radio wave source 20 is located in the fourth divided area 18, the RSSI value obtained by the fourth sensor 38 located in the fourth divided area 18 (closest to the fourth divided area 18) often occurs as a high numerical value. Furthermore, the RSSI value obtained by the first sensor 32 located in the first divided area 12, which is farthest from the fourth divided area 18, often occurs as a low numerical value. Furthermore, it can be seen that the RSSI values ​​obtained by the second sensor 34 and the third sensor 36 located in the second divided area 14 and the third divided area 16, which are intermediate positions between the fourth divided area 18 and the first divided area 12, often occur as intermediate values ​​between the RSSI value of the fourth sensor 38 and the RSSI value of the first sensor 32.

[0054] 9 to 12, the histograms of the occurrence probability of the RSSI values ​​of each sensor 30 are converted into a Gaussian distribution, but they may be converted into other probability distributions or left as histograms. As described above, the advance preparation step of this embodiment involves dividing the data transmission target area 10 into a plurality of divided areas and tallying up the probability distribution of the RSSI values ​​of each sensor 30 for radio waves generated from the radio wave source 20 in each divided area.

[0055] 9 to 12 correspond to the probability distribution of occurrence of sensor observation data by divided area, which is made up of the first mapping data 62, second mapping data 64, third mapping data 66, and fourth mapping data 68 of the RSSI values ​​of each sensor 30 for radio waves generated from the radio wave source 20 in each divided area. By obtaining such a probability distribution of the RSSI values ​​of each sensor 30 for radio waves generated from the radio wave source 20 in each divided area, it is possible to estimate the position of the radio wave source 20 (in which divided area the radio wave source 20 is located) based on the RSSI value of each sensor 30 in the data transmission target area 10.

[0056] If the position of the radio wave source 20 in the data transmission target area 10 can be identified in this manner, it is possible to grasp in more detail the tendency of the probability of the RSSI value that each sensor 30 can have. After the data transmission target area 10 is divided into a plurality of divided areas in this manner, the first mapping data 62 to fourth mapping data 68 of the RSSI values ​​of each sensor 30 shown in Figures 9 to 12 are calculated, and then the process proceeds to the next processing step.

[0057] Specifically, a system designer compares the RSSI value state of each sensor 30 that receives radio waves from a certain radio wave source 20 in a data transmission target area 10 with the first mapping data 62 to fourth mapping data 68 of the RSSI values ​​of each sensor 30 shown in FIGS. 9 to 12 in a state in which the data transmission target area 10 is divided into a plurality of divided areas, which was compiled in a preliminary preparation step. The system designer then selects any one of the first mapping data 62 to fourth mapping data 68 that has the highest similarity, and performs optimization design of an index to be used in the PLIM method to be used in the data transmission target area 10 using the selected one of the first mapping data 62 to fourth mapping data 68. Specifically, as in the reference embodiment, the first embodiment, and the second embodiment, it is sufficient to calculate solutions to the equations in Equation 3, Equation 4, and Equation 5, which are equations for solving a quadratic integer programming problem.

[0058] As described above, according to the data transmission method of this embodiment, the mapping data 60 can be created in detail, and the cumulative error when using replacement data in the event of packet collision at the aggregation station 40 can be further reduced.

[0059] Although the present invention has been described above based on the embodiments, the present invention is not limited to the above embodiments. For example, in the above embodiments, the data transmission method using the data transmission system 50 in which the observation target is the radio wave source 20 is described, but the observation target is not limited to the radio waves transmitted from the radio wave source 20. Other examples of observation targets include the amount of heat from a heat source, the water level from a water source, the wind force from a wind source, and the seismic intensity from a seismic source.

[0060] Furthermore, in the above embodiment, as shown in FIG. 1, an example is given in which the aggregation station 40 is installed outside the range of the data transmission target area 10, but it is also possible to adopt a configuration in which the aggregation station 40 is installed within the range of the data transmission target area 10.

[0061] Furthermore, in the above embodiment, the observation data to be transmitted is a quantized data number obtained by quantizing the RSSI value of the reception strength of the radio wave received by the sensor 30 from the radio wave source 20 within a required numerical range, but the observation data transmitted from the sensor 30 to the aggregation station 40 is not limited to this form. The observation data received by each sensor 30 may be transmitted directly from each sensor 30 to the aggregation station 40.

[0062] Furthermore, the configuration of the present embodiment described above may be appropriately combined with any of the modified examples described in the specification or other known configurations. [Explanation of symbols]

[0063] 10: Data transmission area 12: First division area, 14: Second division area, 16: Third division area, 18: 4th division area 20: Radio source (observation target) 30: Sensor 32: First sensor, 34: Second sensor, 36: Third sensor, 38: Fourth sensor 40: Aggregation station 50: Data transmission system 60: Mapping data (sensor observation data occurrence probability distribution) 62: First mapping data (probability distribution of sensor observation data by divided area), 64: Second mapping data (probability distribution of sensor observation data by divided area), 66: Third mapping data (probability distribution of sensor observation data by divided area), 68: 4th mapping data (probability distribution of sensor observation data by divided area)

Claims

1. A data transmission method for transmitting observation data to a central station by a PLIM (Packet Level Index Modulation) method in a data transmission area in which a plurality of sensors are installed to observe an observation target, the observation data from each of the sensors is packetized, and the packetized observation data is assigned to any of indexes corresponding to frames formed by a plurality of time slots and channels in use, the method comprising: a preparation step of aggregating the observation probability of the numerical value of the observation data at each of the sensors and the position of the observation target as a sensor observation data occurrence probability distribution; calculating the index that minimizes a cumulative total of errors between packet data received at the aggregation station in the PLIM scheme used by each of the sensors and replacement data when the packet data cannot be received at the aggregation station, using the sensor observation data occurrence probability distribution based on the observation data by each of the sensors within the data transmission target area; a step of each of the sensors transmitting the observation data to the aggregation station using the PLIM method to which the index is applied; and the observation data is quantized by a plurality of observation data numbers corresponding to each required numerical range, a step of calculating an index that minimizes a cumulative error between the packet data received at the aggregation station and the replacement data when the packet data cannot be received at the aggregation station in the PLIM method used by each of the sensors, the step being performed based on the following equation: [Equation 6]

2. When the observation data is an average value of the observation data of each of the sensors calculated in the advance preparation step, Each of the sensors omits transmitting the packet data for the observation data to the aggregation station; 2. The data transmission method according to claim 1, wherein the central station adopts the average value as the replacement data when the packet data cannot be received.

3. A data transmission method for transmitting observation data to a central station by a PLIM (Packet Level Index Modulation) method in a data transmission area in which a plurality of sensors are installed to observe an observation target, the observation data from each of the sensors is packetized, and the packetized observation data is assigned to any of indexes corresponding to frames formed by a plurality of time slots and channels in use, the method comprising: a preparation step of aggregating the observation probability of the numerical value of the observation data at each of the sensors and the position of the observation target as a sensor observation data occurrence probability distribution; calculating the index that minimizes a cumulative total of errors between packet data received at the aggregation station in the PLIM scheme used by each of the sensors and replacement data when the packet data cannot be received at the aggregation station, using the sensor observation data occurrence probability distribution based on the observation data by each of the sensors within the data transmission target area; a step of each of the sensors transmitting the observation data to the aggregation station using the PLIM method to which the index is applied; and the observation data is quantized by a plurality of observation data numbers corresponding to each required numerical range, a step of calculating an index that minimizes a cumulative error between the packet data received at the aggregation station and the replacement data when the packet data cannot be received at the aggregation station in the PLIM method used by each of the sensors, the step being performed based on the following equation: [Equation 7]

4. When the observation data is the most frequent value of the observation data of each of the sensors calculated in the advance preparation step, Each of the sensors omits transmitting the packet data for the observation data to the aggregation station; 4. The data transmission method according to claim 3, wherein the central station adopts the most frequent value as the replacement data when the packet data cannot be received.

5. The advance preparation step includes: a step of dividing the data transmission target area into a plurality of divided areas, and aggregating the observation probabilities for the numerical values ​​of the observation data at each of the sensors in a specific divided area as a sensor observation data occurrence probability distribution by divided area for each of the divided areas, The step of calculating the index that minimizes the cumulative total of errors between the packet data received at the aggregation station in the PLIM method used by each of the sensors and the replacement data when the packet data cannot be received at the aggregation station includes: selecting the most similar sensor observation data occurrence probability distribution for each divided area from the sensor observation data occurrence probability distributions for each divided area based on the observation data obtained by each of the sensors within the data transmission target area; The data transmission method according to any one of claims 1 to 4, characterized in that it is a step of calculating the index that minimizes the cumulative total of errors between the packet data received at the aggregation station and the replacement data when the packet data cannot be received at the aggregation station in the PLIM method used by each of the sensors, using the selected probability distribution of occurrence of sensor observation data for each divided area.

Images